an analysis of nest box use by purple martins, house … · 2015. 3. 4. · jerome a. jackson and...

AN ANALYSIS OF NEST BOX USE BY PURPLE MARTINS,

HOUSE SPARROWS, AND STARLINGS IN EASTERN NORTH AMERICA

JEROME A. JACKSON AND JAMES TATE, JR.

In eastern North America the Purple Martin (Progne subis) now nests

almost exclusively in houses provided by man. Recognizing this dependency

on and popularity with man, Jackson initiated a survey of martin populations

in Mississippi in 1971. The purposes of the survey were to monitor population fluctuations of the species over a period of years and to determine the

habitat characteristics that favor a successful martin colony. The survey met

with such overwhelming response from the general public that it was soon

expanded to include all of eastern North America. The questionnaires that

were distributed also asked for information relating to the population biology

of other hole-nesting species, particularly House Sparrows (Passer domesti-

cus) and Starlings (Sturnus vulgar-is), that compete with martins for nest

sites. This paper is an analysis of the response of volunteer cooperators to

two questionnaires that were distributed for the 1971 breeding season. The

previous published surveys of Purple Martin populations by Bartel (1947,

1959)) Mayfield (1964, 1969)) and Hunter (1967) were on a smaller scale

within small geographic areas, and they considered population numbers with-

out regard for nest site characteristics.

THE QUESTIONNAIRE AND ANALYSIS

The initial single page questionnaire included the following questions:

1.

2.

3. 4.

5.

We have (number) martin house(s) and (number) gourds for martins. These pro- vide a total of (number) nest compartments. At the present time (day month year) we have the following actual or approximate number of individuals occupying our nest compartments: Purple Martins (num- her), House Sparrows (number), Starlings (number), Other (Specify) (number). Our martin house and/or gourds are approximately (number) feet above the ground. The area immediately surrounding our martin house and/or gourds (within about 50 ft.) (Circle one) : 1) is a lawn or pasture with few small shrubs; 2) is a lawn or pasture with one or more small trees; 3) is a lawn or pasture with one or more large trees; 4) Other (Specify). Our martin house and/or gourds are approximately (number) feet from the nearest building.

This questionnaire was freely distributed to any potential respondent, and was completed by 1067 persons from 32 states, the District of Columbia, and five Canadian provinces. These areas and numbers of respondents were as follows: Alabama (5),

Arkansas (47), Connecticut (8), Delaware (l), District of Columbia (2), Florida (12))

Georgia (l), Illinois (16), Indiana (3), Iowa (108), Kansas (2), Louisiana (8), Maine

435

THE WILSON BULLETIN December 1974 Vol. 86, No. 4

(12), Maryland (ll), Massachusetts (l), Michigan (111, Minnesota (181, Mississippi (5921, Missouri (41, Nebraska (51, New Hampshire (11, New Jersey (141, New York (771, North Dakota (11, Ohio (lo), Okl h a oma (12)) Pennsylvania (261, Rhode Island

cl), South Carolina (7), Tennessee (9), Texas (51, Virginia (81, Wisconsin (81, Al- berta (41, Manitoba (l), Nova Scotia (11, Ontario (121, Quebec (3).

The phraseology of our question relating to the number of birds using the apartments resulted in some misinterpretation and inconsistent data. We asked for the number of each species using the apartments, so that at times nestlings as well as adults were counted. Where this was apparent, these data were questioned and corrected or dis- carded.

For the purpose of this study, the word colony is used to refer to all of the apartments supplied by one respondent at one locality. A colony may include one or more apartment house and/or gourds. An apartment is defined as a single potential nest site, whether it is a separate unit, such as a gourd, or part of a multiple unit structure (i.e., an apartment house).

Using the data from the first questionnaire, we calculated the ratios of martins, sparrows, and Starlings per apartment. In addition, some comparisons of martin and spar-

row populations and the relative use by the birds of apartments in houses and gourds in Mississippi were possible because of the large number of cooperators from that state. These ratios were compared for different geographic areas, and for differences in apartment heights, distance of apartments from the nearest building, and immediate environ- ment of the apartments.

The second questionnaire, sent to respondents to the first, consisted of three pages relating to apartment characteristics, colony history, habitat, apartment maintenance, and apartment use by the birds. Only data for apartment houses (N = 622) are included here in our analysis of the second questionnaire data. Separate questionnaires were completed for each multiple-apartment house. Questions regarding the characteristics and location of the house included the following:

1. We have Inumber) multiple-apartment houses and (number) single-apartment houses or gourds.

2. The multiple-apartment house here described was first erected on the present site in (rear).

3. There are other martin colonies within one-half mile of this one. 1. Yes 2. No.

4. This house is 1. Aluminum 2. Wood 3. Plastic 4. Fiberglass 5. Ceramic 6. Other.

5. The outside of this house is basically 1. White 2. Green 3. Brown 4. Red 5. Blue 6. Other.

6. The inside of the apartments is basically 1. White 2. Green 3. Brown 4. Other.

7. This house is in the shade 1. Only in early morning 2. Most of the morning 3. Most of the day 4. Most of the afternoon 5. Only in late afternoon 6. All of the day 7. Never in the shade 8. Other.

8. The location of this house can best he described as 1. Urban, business, or indus- trial area 2. Urban, residential area 3. Suburban 4. Rural 5. Other.

9. There is a lake, pond, stream, or river that can be seen from the martin house. 1. Yes 2. No.

10. There are power or telephone lines that can be seen from the martin house. 1. Yes 2. No.

11. We clean the apartments out each year. 1. Yes 2. No 3. Sometimes 4. Does not apply.

NEST BOX ANALYSIS 437

FIG. 1. Arbitrary (and a priori) division of eastern North America into nine sectors for analysis of nest box occupancy. Numbers separated by slashes are the percentages of apartments occupied by Purple Martins, House Sparrows, and Starlings, respectively. Numbers in parentheses indicate the total number of colonies from which data were

received.

12. There are bushes or high shrubbery growing at the base of the pole supporting the house. 1. Yes 2. No.

21. This house is approximately (number) feet above the ground or water.

Data for colonies using gourds were available from too few colonies to permit mean-

ingful statistical analysis. Rather than asking for the number of individuals using the

apartments, on the second questionnaire we asked for the total number of apartments

occupied by each species. Because of the small total sample size the data were grouped

for analysis into nine geographic regions as indicated in Figure 1. These regions will

be referred to in subsequent discussions as the northwest, north-central, northeast, west-

central, central, and so on.

Using data from the second questionnaire, we calculated the percentage of apartments

in each colony that were occupied by martins, sparrows, and Starlings, and compared

variation in these percentages with variation in several physical and biotic parameters.

Statistical comparisons of data from both questionnaires were made by using a simple

analysis of variance, the sum of squares simultaneous test procedure (STP) (Sokal and

Rohlf, 1%9:68s687), the test for equality of two percentages discussed by Sokal and

Rohlf (1969:607-608), or the method of Brandt and Snedecor (Snedecor, 1956:227-230)

for comparison of sets of more than two percentages.

Sample sizes for the various data sets refer to the number of colonies or houses; per-

centages refer to the number of apartments occupied.


RESULTS

First Questionnaire.-One of the major findings from the first question-

naire was a statement of the relative abundance and distribution of Purple

Martins. This information is given in Table 1, expressed as ratios of oc-

cupancy for each state or Canadian province from which there were five or

more cooperators. In general, the number of martins per apartment is highest in southwestern Ontario along Lake Erie, along the Maine coast, and in

the mid-south (Tennessee and Mississippi). The lowest ratios are from parts

of the mid-west, middle Atlantic states, and Florida. An analysis of variance

indicates that there is significant geographic variation in the use of apart-

ments by martins (p < 0.05).

TABLE 1

RATIO OF MARTINS/APARTMENT FOR EACH STATE OR CANADIAN PROVINCE REPRESENTED BY

DATA FROM FIVE OR MORE COLONIES

state or Total number province of apartments

‘Total number of Purple Martins

Ratio: martins/apartment

Ontario 232 363 1.56

Maine 310 360 1.16

Tennessee 314 362 1.15

Mississippi 13133 10797 0.82

Oklahoma 182 133 0.73

Virginia 152 107 0.70

New York 1587 1086 0.68

New Jersey 278 190 0.68

Ohio 237 159 0.67

South Carolina 124 80 0.65

Arkansas 893 561 0.63

Iowa 2592 1620 0.63

Alabama 160 116 0.63

Wisconsin 286 164 0.57

Louisiana 223 121 0.54

Pennsylvania 607 299 0.49

Texas 100 46 0.46

Maryland 217 96 0.44

Illinois 362 151 0.42

Minnesota 418 157 0.38

Nebraska 86 27 0.31

Florida 402 110 0.27

Michigan 174 46 0.26

Connecticut 322 48 0.15

Jackson and T&C3 NEST BOX ANALYSIS

TABLE 2

RATIO OF BIRDS/APARTMENT IN RELATION TO APARTMENT HEIGHT AND VEGETATION

WITHIN 15.2 M OF THE APARTMENTS

Number of colonies

Ratio of birds/apartment

Martins Sparrows Stnrlings

Apartment height (m)

2<3 3<4 4<5 5<6 6<7 7<8

72 0.75 0.20 0.05

202 0.64 0.18 0.01

370 0.67 0.20 0.03

95 0.73 0.23 0.02

125 0.63 0.21 0.03

53 0.73 0.20 0.03

Vegetation

A few small shrubs 187 0.74 0.15 0.03

One cc m0re small trees 366 0.68 0.20 0.03

One or more large trees 532 0.67 0.20 0.03

From that survey we also found the height of the apartments seems to have

little effect on occupancy. Data for apartment height were grouped into six

classes (Table 2), among which no significant differences were found in oc-

cupancy values for any of the species considered. On the other hand, the distance of apartments from the nearest building

seems to be important to martins, but less so to sparrows or Starlings. There

were no significant differences among mean ratios of martins per apartment

for distance subsets closer than 30 m from the nearest building. Apartments

closer than 30 m, however, had significantly more (p < 0.05) martins (mean

= 0.73 martins/apartment) than did apartments that were more than 30 m

distant (mean = 0.55 martins/apartment).

Finally no significant differences were indicated in occupancy rates as re-

lated to habitat types in the vicinity of the apartments. We did note a trend

toward fewer martins and more sparrows (p < 0.1) with increasing height

of vegetation (Table 2)) although a larger sample size and/or a refinement of the question is needed to statistically confirm the relationships.

The very large body of data from Mississippi makes it possible to compare

colonies using gourds, houses, or both. From that state 592 cooperators re-

ported 3,173 gourds and 1,046 houses with a total of 13,133 apartments avail-

able for birds. Totals of 10,797 Purple Martins, 1,754 House Sparrows, and

379 Starlings were reported nesting in these apartments. Table 3 summarizes

the ratios of martins and sparrows per apartment for the three colony types

(Starlings were not analyzed because of their relatively smaller number).


TABLE 3

RATIO OF MARTINS/APARTMENT AND SPARROWS/APARTMENT IN GOURDS AS

OPPOSED TO HOUSES IN MISSISSIPPI

Colony type Number of colonies Martins/apartment Sparrows/apartment

Gourds only 73 1.20 0.07

Houses only 352 0.83 0.17 Both houses and gourds 164 0.66 0.18

The three ratios for martins all are significantly different (p < 0.01). The

ratio of sparrows/apartment for gourd colonies is significantly less than the

ratio for either colonies with houses or colonies with both houses and gourds

(p < 0.02). Comparison of the ratio of martins/apartment in house and gourd colonies

in Mississippi with and without sparrows also indicates significant differences

(p < 0.05). In colonies with apartment houses and no sparrows there were

more martins per apartment (ratio = 1.02, n = 122) than in colonies without

sparrows (ratio = 0.79, n = 217). Th is was also true in colonies with gourds

as apartments (no sparrows: ratio = 1.36, n = 49; with sparrows: ratio =

0.98, n = 21). There were also significantly more martins in gourd colonies

without sparrows than in house colonies without sparrows. The difference between house and gourd colonies with sparrows was not significant, nor was

that between house colonies without sparrows and gourd colonies with spar-

rows.

Second Questionnaire.-Analysis of data from the second questionnaire

provides more insight into the geographic variation in relative population

levels and the factors influencing nest-site selection in the three species.

Analysis of the percentages of apartments used by the birds in the nine

geographic sectors (Fig. 1) confirms the trends indicated by data from the

first questionnaire. STP analysis reveals two non-significant subsets based

on the ranked means of percentages of apartments used by martins. Oc- cupancy is higher in the south-central than in the central region, but differ-

ences between other pairs of sectors are not statistically significant.

Significant geographic variation is also indicated in the relative use of

apartments by sparrows. Two non-significant subsets were defined by STP

analysis-the central sector had a significantly higher proportion of spar-

rows than any other area. There were no significant differences between any

pair of the remaining sectors.

Analyses of the types of house (Table 4) showed higher percentage of oc-

cupany in wooden than in alumninum houses for all three species; however,

Jackson and T&S NEST BOX ANALYSIS 441

TABLE 4

PERCENTAGE OF APARTMENTS OCCUPIED RELATIVE TO THE PHYSICAL CHARACTERISTICS

AND GENERAL LOCATION OF THE HOUSE

Number of Percentage of apartments occupied by:

HOUSf3 Apartments Martins Sparrows Starlings

Type of house

Aluminum Wood

Exterior color

White Brown Green

Interior color White Brown

Apartments clean in Spring

Yes No

Location

Urban-business Urban-residential Suburban Rural

Water within sight

Yes No

Neighboring martin house

Yes No

Utility wires within sight

Yes No

Shrubs at base of support pole

Yes

No

224 2946 40.6 11.0 0.3 394 5379 44.1 12.4 1.8

556 20 28

11.7 1.0 16.1 8.3 17.9 2.6

295 282

7585 43.9 218 28.4 341 30.8

3798 43.2 3886 44.9

10.2 1.0 13.9 1.6

526 7200 41.9 12.5 0.9 43 502 53.4 9.6 3.4

22 314 57.3 13.4 0.3 253 3204 45.1 13.1 1.2

147 2036 34.6 14.4 1.7

198 2791 45.1 8.6 1.2

295 3920 45.6 10.5 1.4 324 44Q9 48.5 13.1 1.2

498 6569 43.9 13.0 1.1 121 1748 39.5 8.0 1.8

584 7843 42.6 12.0 36 502 47.4 10.4

137 1910 43.2 14.8

481 6401 42.8 11.1

1.3 0.4

1.7

1.1


TABLE 5

PERCENTAGE OF APARTMENTS OCCUPIED IN RELATION TO THE AMOUNT OF TIME A HOUSE Is

IN THE SHADE

Number of

HOUSeS Apartments

Percentage of apartments occupied by:

Martins SpkVKXVS Starlings

Time during which the house is shaded

Never 353 Always 13 Only late afternoon 108 Most of the

afternoon 10 Most of the day 12 Only early morning 99 Most of the morning 20

4934 43.5 12.2 1.4 182 46.2 9.3 0.0

1466 45.4 9.4 1.4

108 43.5 19.4 0.9 150 36.0 24.7 5.3

1224 40.1 12.2 0.7 211 32.2 11.4 0.0

the difference approaches significance only for Starlings (p G 0.01). Most

houses (92 percent) were painted white; thus, in spite of large differ-

ences in the average percent occupancy, sample sizes for brown and green

houses are too small to prove the differences are statistically significant. The

data do suggest that martins might favor white houses (p < 0.2) while Star-

lings favor darker ones (p < 0.01). Martins and Starlings showed no clear

preference for white or dark (brown or natural wood) interiors in their apart-

ments, but sparrows may prefer (p < 0.2) the darker colors.

Comparison of houses that were cleaned out before the beginning of the

TABLE 6

PERCENTAGE OF APARTMENTS OCCUPIED IN 1971 AS IT RELATES TO TIIE YEAR IN WHICH A

HOUSE WAS FIRST ERECTED

Year house was first erected

1971

1970 1969 1968 1967 1966 1965

Before 1965


HOUSeS Apartments Martins SparrOWS Starlings

113 1398 30.0 10.0 1.4

86 1165 43.0 9.7 1.1 90 1125 48.4 11.3 1.3

91 1160 42.5 12.8 1.5 53 677 43.6 15.2 1.9 46 608 50.5 14.8 1.6 40 555 56.6 11.9 0.0 93 1558 41.5 12.9 1.0

Jackson and T&e


TABLE 7

TIIE PERCENTAGE OF APARTMENTS OCCUPIED IN RELATION TO THE NUMBER OF APART-

MENTS/HOUSE


HLXW3 Apartments Martins Sparrows Starlings

Number of apartments/house

8 or fewer 9-12

13-16 17-20 21-24 2Fk28

122 789 49.9 14.6 1.8 280 3284 41.9 11.6 0.9

91 1403 43.6 15.0 1.5 68 1273 47.8 8.9 2.1 46 1132 32.8 11.1 0.7

3 80 52.5 2.5 0.0

breeding season and those that were not suggests that martins might prefer

to use an old nest as a base for a new one (p < 0.2). Only 43 cooperators

failed to clean out their bird houses.

Analyses of the general location of houses (Table 4) show that suburban colonies had significantly fewer martins (p < 0.05) than did urban-business, urban-residential, or rural colonies. Rural martin houses tended to have fewer sparrows than did either suburban (p < 0.1) or urban-residential colo-

TABLE 8

PERCENTAGE OF APARTMENTS OCCUPIED BY SPECIFIC COMBINATIONS OF MARTINS,

SPARROWS, AND STARLINGS


Birds Present HOUSfX Apartments Martins Sparrows Starlings

No birds 53 769 0.0 0.0 0.0

Martins only 205 2575 57.1 0.0 0.0

Sparrows only 31 382 0.0 20.2 0.0

Starlings only 6 88 0.0 0.0 20.5

Martins and

sparrows only 281 3836 48.1 20.4 0.0 Martins and

Starlings only 9 164 50.0 0.0 13.4

Sparrows and

Starlings only 13 183 0.0 33.9 16.9

Martins, sparrows, and

Starlings 24 375 50.1 19.2 9.1

444 THE WILSON BULLETIN December 1974 Vol. 86. No. 4

nies. The presence of a pond, lake, or stream within sight of a house may

result in a greater occupancy by martins (p < 0.2). Occupancy by sparrows

was less, but not significantly so. The occurrence of another house within

one-half mile had no significant effect on the percent occupancy by mar-

tins, but the increase in the percent occupany by sparrows approaches sig-

nificance (p < 0.15). The presence of utility wires within sight had no dis-

cernible effect on the percent occupancy of a house by any of the species. The presence of bushes or other high vegetation at the base of the support

pole apparently does not influence the rate of occupancy of the house by mar-

tins, but may favor occupancy by sparrows (p < 0.25). Statistical compari-

son of colonies that are in the shade at various times (Table 5) indicates no

significant differences for any species, though some trends are suggested.

Martins may prefer houses that receive morning sun, and sparrows seem to

prefer houses that receive some but not afternoon sun. Larger sample sizes

are needed to confirm these patterns.

The percent occupancy by martins for houses erected in 1971 (Table 6)

is significantly lower (p < 0.05) than the average percent occupancy for nest boxes erected earlier. No such difference is apparent for sparrows or

Starlings. Comparison of occupancy rates in relation to the number of apart-

ments in a house (Table 7) shows no immediate pattern, though the percent-

age of apartments occupied by martins in houses with eight or fewer apart-

ments is significantly greater than the value for houses with nine to 12 apartments.

Finally, a comparison of houses only occupied by specific combinations of

these three species (Table 8) indicates that houses with only martins have

significantly more martins than houses that also have sparrows. There were

no significant differences within the sets of combinations including sparrows or Starlings.

DISCUSSION Nest Site Preferences.-Attempts to manage game, aesthetically pleasing,

or rare species have multiplied in recent years with the advent of a greater

ecological awareness. The Purple M ar in, t however, is a species that was

managed in North America even before the arrival of European man. Indians

in the Southeast hung calabash gourds from trees near their villages or from

poles in their corn fields. Martins accepted these nest sites and, as a result of

their mobbing behavior, kept predators away from the corn and foodstuffs

that were hung up to dry (Audubon, 1831; Ewan and Ewan, 1970). The

early colonists substituted ceramic jugs and wooden apartment houses for

gourds in attempts to attract martins. For example, Audubon (op. cit.) noted

that “Almost every country tavern has a martin box on the upper part of its sign-board.”

Jackson and T&e


Through the years, martin houses have appeared in as many styles, shapes,

and sizes as human dwellings, and a lot of generalities have been made about

what will and what won’t attract martins. While man has “managed” mar-

tins for centuries, the management has been based primarily on tradition, and

testing of management practices has been limited to trial and error. One re-

sult of our study is that the putative nest site requirements of martins have

been objectively analyzed. Such “requirements” have been enumerated many

times (e.g., Wade, 1966; Layton, 1969) and generally include such things as:

1, the apartment house should be in an open area at least 8 m from the near-

est tree or building; 2, the house should be on a pole 5-10 m above the

ground; 3, martins need or prefer utility wires near the house; 4, a pond,

lake, or stream nearby is characteristic of a preferred habitat; and 5, the

apartments should be cleaned out each year.

Clark (1970) quantitatively examined the relationship between apartment

characteristics and rate of occupancy by martins, but, because of her small

sample size, her results are inconclusive. Our study confirms some of the

suggested requirements, but refutes others. Our data do not support the

contention that a martin house should be away from trees or buildings. In

fact, houses within 30 m of a building seem to be preferred to those farther

away. No clear preferences are shown with regard to apartment height or the

presence or absence of utility wires. The contention that martins prefer water

nearby is supported.

Most noteworthy is the suggestion that martins prefer apartments that

haven’t been cleaned out before the beginning of the nesting season. Kinney

(1972) notes the use of old nests by Purple Martins, and one of our coopera-

tors reported that the first apartments occupied by martins are always those

containing old nests. Boyd (1935) points out that House Martins (Delichon

urbica) “generally reline and use an old nest,” and frequent use of old nests

has been observed for the Cliff Swallow (Petrochelidon pyrrhonota) (BUSS,

1942; Samuel, 1971) and Barn Swallow (Hirundo rustica) (Samuel, 1971).

The potential energetic savings resulting from use of an old nest is obvious

and might be expected to be selected for, though an inherent disadvantage is

that old nests harbor parasites that may reduce the fecundity of the birds

(Moss and Camin, 1970). From a Purple Martin management point of view

we recommend that House Sparrow and Starling nests be removed and that

martin nests be left, but dusted with some non-persistent miticide.

It appears that we have been trying to attract martins by following a set

of guidelines that for the most part either do not influence the birds or that

influence them negatively. Considering this revelation, it would seem ap-

propriate that other management programs, particularly those involving en-

446 THE WILSON BULLETIN December 1974 Vol. 86, No. 4

dangered species, be examined for practices that are based on tradition rather

than on objective analysis of data. First Year Occupancy of Houses by Martins.-Mayfield (1964, 1969) in a

fifteen-year study of martin houses in Ohio, reported the houses as fully oc-

cupied their first year as in subsequent years. However, a reexamination of

his data suggests that his sample sizes may have been too small to give sta-

tistically valid results. For the first ten years of the study (Mayfield, 1964)

18 of 32 houses had fewer martins the first year than the average for all years.

A comparison of this number with the number of houses having the average

or a greater number of martins indicates no significant differences (x2 =

0.50). With five more years data (Mayfield, 1969) the differences are still

not significant (21 houses with fewer than the average number of martins the

first year, 15 with the average or more; x 2 = 1.0)) but the increased chi square value and trend of the data suggest that only increased sample size is needed

for statistical proof. In our study comparison of a large number of houses

for a single year confirms the traditional suggestion that first year houses are

likely to attract fewer martins. This may he because houses are frequently

placed by inexperienced people on a trial-and-error basis and often moved

if unused. Such a difference probably also reflects the tendency for young

martins to return to the general area of the colony in which they were hatched

and for adults to return to the same colony to nest year after year (Allen and

Nice, 1952; Olmstead, 1955).

Competition Among Species for Nest Sites.-The data presented here sug-

gest that Starlings are not a serious threat to the Purple Martin as a nest site

competitor (though they may be a serious threat as a predator on nestlings

and eggs (Flentge, 1940; Gaunt, 1959) ) . Kessel (1957) never found more than one pair of Starlings nesting in a multi-apartment house. She states that

one male will defend an entire apartment house against other Starlings but

that “once settled in his small domain . . . the starling becomes an unob-

trusive neighbor” to other species. Finally, as Wade (1966) suggests and

this study confirms, aluminum houses which are used widely for martins,

further deter the Starling from competing with martins. House Sparrows, unlike Starlings, will nest communally and therefore

present more of a problem for martins. The relative degree to which these

species have adapted to communal nesting is perhaps reflected by the average

percentage of apartments they occupy in a colony. We found these figures to

be 48.1 percent for martins and 20.4 percent for sparrows, and Olmstead

(1955) reported similar values (50.0 percent for martins, 18.2 percent for

sparrows) for a Kansas colony he studied for five years. With such a high

percentage of apartments occupied by sparrows, it is not surprising that the

presence of sparrows results in a significant decrease in the number of apart-


ments martins occupy. The effect of sparrows, however, may be more than

a proportionate loss of potential martin nest sites. Pindar (1923) suggests

that if there are only a few martins at an apartment house they may be com-

pletely repelled by sparrows. Furthermore, when there are a large number of martins such that most apartments are occupied, the male House Spar-

rows are likely to usurp the martins’ nests and destroy their eggs (Olmstead,

1955). Geographic variation in the relative abundance of martins, sparrows, and

Starlings may be related to interspecific competition. In the central region

of the eastern United States (Fig. 1)) martins are at their lowest, sparrows

at their highest, and Starlings at a low relative abundance. In the northeast

and east-central regions martins are at a low, sparrows at their lowest, and

Starlings at their highest levels. In the south-central region martins are at

their highest, and sparrows and Starlings are both at low levels. Levels of all

three species may be reduced in the northern sections as a result of some competition with Tree Swallows (Zridoprocne bicolor), which are known to

nest in martin houses (Tate, 1963; observations reported by participants in

this study).

Additional limiting factors may contribute to the regional differences in

relative abundance of the species. Climate during the breeding season (hence

supply of flying insects) is a known limiting factor for martins (Horton,

1903; Tate, 1972). Amount of land under cultivation for grain crops or

number of head of granivorous livestock (hence food supply) may limit spar-

rows. Suggestive of this is the fact that the central region, where sparrows are relatively most numerous, is well defined (Jones, 1972) as the major

area of corn and hog production in North America. Time may be the limit-

ing factor for Starlings. Dispersal of the Starling into areas outside the

northeast and east-central regions has been since 1926 (Kalmbach, 1928),

and insufficient time may have lapsed for breeding populations to build up

to levels similar to those in the northeast. If time is the limiting factor, both

martin and sparrow populations can be expected to decrease as the Starlings

become more established.

FUTURE WORK

The Purple Martin survey has been continued and enlarged since 1971 in

order to monitor population fluctuations. It is pure serendipity that in 1972,

after population levels had been documented for one year, Purple Martin pop-

ulations were decimated by Hurricane Agnes (see Tate, 1972). Thus, with

knowledge of previous levels, we can quantify the effects of the storm on

martins as well as their reestablishment. Other work stemming from the sur-

vey will include multivariate analysis of the data sets presented here and


behavioral and ecological studies of nest site selection and interspecific com-

petition. Persons interested in participating in the survey are urged to con-

tact the senior author.

SUMMARY

A public response survey seeking information about Purple Martin, House Sparrow, and Starling populations and nest site characteristics reveals that: 1, significant geographic variation occurs in the rate of occupancy of apartment houses by Purple Martins and House Sparrows; 2, height of apartments has no significant effect on the rate of occupancy by any of the species; 3, apartments farther than 30 m from the nearest building had significantly fewer martins than closer apartments; 4, there may be fewer martins and more sparrows occupying apartments as height of vegetation in- creases; 5, in Mississippi, martins show a significant preference for and sparrows a significant avoidance of gourds as nest sites; 6, Starlings may avoid aluminum houses; 7, martins may prefer white apartments while Starlings may favor darker ones; 8, martins tend to prefer apartments that have not been cleaned out after a previous occupancy; 9, suburban have significantly fewer martins than do urban or rural colonies and rural tend to have fewer sparrows than do suburban or urban colonies; 10, the presence of a pond, lake or stream within sight of the apartments may result in a greater occupancy by martins; 11, the presence of other apartment houses within one-half mile has no significant effect on occupancy by martins, but may result in a greater occupancy by sparrows; 12, significantly fewer apartments are occupied the first year by martins than in subsequent years; 13, competition between martins and sparrows may result in significantly fewer martins occupying a colony, but the Starling is not a serious nest-site

competitor.

ACKNOWLEDGMENTS

Our first thanks must go to the cooperators who participated in this survey. Without their help this study would not have been possible. Second, we are further indebted to the many cooperators who contributed financially to the program and to the Mississippi State University Development Foundation for managing their contributions. Generous financial assistance was also provided by Mississippi State University and by the Cornell Laboratory of Ornithology. Finally, a special thanks is due to Connie Jones and Ken- neth Bicker who assisted in preparing the questionnaires for mailing and the returned

data for processing.

LITERATURE CITED

ALLEN, R. W., AND M. M. NICE. 1952. A study of the breeding biology of the Purple Martin (Progne subis). Amer. Midl. Nat., 47:60&665.

AUDUBON, J. J. 1831. Ornithological biography. Vol. 1. Philadelphia. BARTEL, K. E. 1947. Increase of Purple Martins. Audubon Bull., 61:5-6. BARTEL, K. E. 1959. Purple Martin fluctuations. Inland Bird Banding News. 31(6) :

63.

BOYD, A. W. 1935. Report on the Swallow enquiry, 1934. British Birds, 29:3-21.

Buss, I. 0. 1942. A managed Cliff Swallow colony in southern Wisconsin. Wilson

Bull., 543153161.

Jackson and Tate NEST BOX ANALYSIS 449

CLARK, M. C. 1970. Purple Martin houses and factors which may determine their being occupied. Master’s Thesis, Pennsylvania State University.

EWAN, .I. AND N. EWAN. 1970. John Bannister and his natural history of Virginia 1678- 1692. Univ. Illinois Press, Urbana.

FLENTGE, L. G. 1941). The European Starling-friend or foe? Audubon Bull., 1940 (34) :1-6.

GAUNT, A. S. 1959. Behavior in the Purple Martin. Kansas Omithol. Sot. Bull., 10: 14-16.

HORTON, F. B. 1903. Mortality of Purple Martins (Progne purpurea) at Brattleboro, Vt. Auk, 20:435436.

HUNTER, R. E. 1967. Purple Martin survey in New Brunswick. Moncton Publishing Co., Ltd., Moncton, N. B.

JONES, D. B. ted.). 1972. Oxford economic atlas of the world, 4th ed. Oxford Univer- sity Press, London.

KALMBACH, E. R. 1928. The European Starling in the United States. U.S. Dept. Agric., Farmers’ Bull., No. 1571.

KESSEL, B. 1957. A study of the breeding biology of the European Starling @turnus vulgaris L.) in North America. Amer. Midl. Nat., 58:257-331.

KINNEY, F. W. 1972. Who needs Purple Martin birds. Rowley Printing, Inc., Rowley, Mass.

LAYTON, R. B. 1969. The Purple Martin. Nature Books Publishers, Jackson, Miss. MAYFIELD, H. F. 1964. Yearly fluctuations in a population of Purple Martins. Auk,

81:274-280. MAYFIELD, H. F. 1969. Purple Martin population changes over fifteen years. Auk, 86:

522-528. MOSS, W. W., AND J. H. CAMIN. 1970. Nest parasitism, productivity, and clutch size

in Purple Martins. Science, 168: 1000-1003. OLMSTEAD, R. 1955. Observations on Purple Martins. Kansas Ornithol. Sot. Bull., 6:

8-10. PINDAR, L. 0. 1923. Some miscellaneous notes on birds. Wilson Bull., 35:159-164. SAMUEL, D. E. 1971. The breeding biology of Barn and Cliff Swallows in West Vir-

ginia. Wilson Bull., 83:284+301. SNEDECOR, G. W. 1956. Statistical methods. 5th ed. Iowa State College Press, Ames. SOKAL, R. R., AND F. J. ROHLF. 1969. Biometry. W. H. Freeman and Company, San

Francisco. TATE, J., JR. 1963. Tree Swallow nesting in martin colony. Wilson Bull., 75:278. TATE, J. L., JR. 1972. The changing season. Amer. Birds, 26:828-831. WADE, J. L. 1966. What you should know about the Purple Martin. J. L. Wade,

Griggsville, Illinois.

DEPARTMENT OF ZOOLOGY, MISSISSIPPI STATE UNIVERSITY, MISSISSIPPI STATE,

MISSISSIPPI 39762 AR’D CORNELL LABORATORY OF ORNITHOLOGY, ITHACA,

NEW YORK 14850 (PRESENT ADDRESS JT, JR.: P.O. BOX 2043, DENVER,

COLORADO 80201). ACCEPTED 19 MARCH 1974.

an analysis of nest box use by purple martins, house … · 2015. 3. 4. · jerome a. jackson and...

Documents